SEARCH

SEARCH BY CITATION

Keywords:

  • genitalia;
  • gerridae;
  • mating success;
  • selection gradients;
  • sexual conflict;
  • sexual selection

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Study animal
  6. Sampling and measurements
  7. Measures of selection, statistical analyses
  8. Results
  9. Net sexual selection
  10. Direct sexual selection
  11. Relation between patterns of sexual selection on total length and genital length
  12. Discussion
  13. Sexual selection on male genital length and its relation to the selection on total length
  14. Sexual selection on the genital components, and the role of precopulatory sexual selection in shaping male genitalia
  15. Evolution of male genitalia: conclusions and perspectives
  16. Acknowledgments
  17. References

While congruent evidence indicates that sexual selection is the most likely selective force explaining the rapid divergence of male genital morphology in insects, the mechanisms involved in this process remain unclear. In particular, little attention has been paid to precopulatory sexual selection. We examine sexual selection for mating success on male genital components in six populations of Aquarius remigis, a water strider characterized by unique genital morphology. Multivariate selection analysis confirms previous findings that precopulatory sexual selection favours longer external genitalia, and provides new evidence that this selection acts independently on external genital components. In contrast, the size of the major internal genital sclerite is not correlated with mating success. Thus, precopulatory sexual selection acts strongly on the size of the external genitalia, but not on the intromittent organ itself. These results highlight the multiple functions of genital organs and the importance of both precopulatory and post-copulatory sexual selection in shaping the remarkable diversity of male genitalia in insects.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Study animal
  6. Sampling and measurements
  7. Measures of selection, statistical analyses
  8. Results
  9. Net sexual selection
  10. Direct sexual selection
  11. Relation between patterns of sexual selection on total length and genital length
  12. Discussion
  13. Sexual selection on male genital length and its relation to the selection on total length
  14. Sexual selection on the genital components, and the role of precopulatory sexual selection in shaping male genitalia
  15. Evolution of male genitalia: conclusions and perspectives
  16. Acknowledgments
  17. References

When developing his theory of sexual selection, Darwin (1871) was concerned with the evolution of sexual dimorphisms and he distinguished between primary and secondary sexual characters. He defined the former as fundamental differences between the reproductive organs of males and females, and attributed their evolution solely to the action of natural selection. While Darwin's distinction has been criticized as not being applicable to many species of plants (Grant, 1995), it has been widely applied to sexually dimorphic traits in animals. However, the distinction between primary and secondary sexual characters of animals is also currently challenged by recent evidence of the importance of sexual selection in the evolution of male genitalia (for review see Cordoba-Aguilar, 2000; Sirot, 2003; Hosken & Stockley, 2004). In promiscuous species with internal fertilization, male genitalia are often complex structures displaying high diversity in their morphology, even within clades of closely related species, and several lines of evidence suggest that sexual selection has been responsible for this rapid divergence (Hosken & Stockley, 2004). This assertion is particularly well supported in insects by comparative studies, detailed descriptions of genital structures and observations of mating interactions (Eberhard, 1985; Arnqvist, 1998; Arnaud et al., 2001; Artiss, 2001). In addition, a growing number of intra-specific studies have revealed correlations between fertilization success and male genital morphology (Arnqvist & Danielsson, 1999; Danielsson & Askenmo, 1999; House & Simmons, 2003; Takami, 2003). Several hypotheses have been proposed to explain this relationship, including sperm competition, cryptic female choice and sexual conflict over fertilization (Simmons, 2001). While the relative importance of these post-copulatory mechanisms in the evolution of male genitalia remains uncertain (Simmons, 2001; Hosken & Stockley, 2004), they have all received considerable theoretical and empirical attention (Simmons, 2001). In contrast, few studies have investigated the influence of male genital morphology on mating success (Simmons, 2001 but see Mason, 1980; Preziosi & Fairbairn, 1996,2000; Arnqvist et al., 1997; Ferguson & Fairbairn, 2000; Sih et al., 2002) in spite of the fact that various models of sexual selection actually predict an association between these two components (Simmons, 2001). Indeed, the evolution of genitalia can be driven by male–male competition for access to females if these structures influence males’ ability either to achieve intromission (scramble competition) or to prevent take-over from other males (contest competition) (Simmons, 2001). Aspects of the genitalia may also serve as sexual armaments, enabling males to overcome female reluctance to mate (Arnqvist & Rowe, 1995,2002a,b). The probable existence of a relationship between genital morphology and mating success is also supported by the reported mechanical functions of male genitalia. Indeed, in a recent review conducted on Diptera, Eberhard (2004a) established that functions associated with intromission are commonly documented. This finding suggests that selection for achieving and/or maintaining intromission is widespread in insects and can play an important role in shaping male genitalia. Empirical estimates of contemporary patterns of selection on male genitalia acting through differential mating success are necessary to test this hypothesis and to improve our current understanding of how sexual selection is driving the evolution of male genitalia.

The water strider Aquarius remigis (Heteroptera; Gerridae) is an excellent model species for such studies. Indeed, species within the Gerridae exhibit very diversified male genital morphologies and the remigis group, in particular, presents a series of pregenital and genital apomorphies (Andersen, 1990; Damgaard et al., 2000; Fairbairn et al., 2003; Damgaard & Cognato, 2003). However, our current knowledge of the adaptive significance of these characteristics is still imperfect (Fairbairn et al., 2003). In A. remigis, mating success significantly influences paternity success (Vermette & Fairbairn, 2002), which indicates that precopulatory mechanisms of sexual selection could have important evolutionary consequences for male morphology. The role of such selective forces in the evolution of male genitalia is substantiated by several studies demonstrating a positive correlation between length of the external genitalia and mating success in male A. remigis (Preziosi & Fairbairn, 1996,2000; Ferguson & Fairbairn, 2000; Sih et al., 2002). This pattern of sexual selection for increased genital length is supported by analyses of selection both in natural populations (Preziosi & Fairbairn, 1996,2000; Ferguson & Fairbairn, 2000) and under controlled conditions in artificial streams (Sih et al., 2002). The mating advantage of males having long genitalia has been ascribed to a better ability to overcome female reluctance to mate (Preziosi & Fairbairn, 1996,2000; Sih et al., 2002; Fairbairn et al., 2003). In A. remigis, mating interactions usually begin with a vigorous premating struggle, the females actively resisting males’ attempts to mate (Weigensberg & Fairbairn, 1994,1996; Lauer et al., 1996; Watson et al., 1998). Several lines of evidence suggest that the elongated genitalia of males, which represent about 20% of total length (Fig. 1a and see Preziosi & Fairbairn, 1996) have evolved in response to the intersexual conflict over mating. In particular, behavioural studies indicate that the mating advantage of males with long genitalia is only due to their higher likelihood of success when attempting to mate (Sih et al., 2002). In addition, close examination of male genital structures and the investigation of their position during the mating interactions indicate that the increased length of various components of male genitalia could serve to facilitate the intromission in spite of the resistance of the female (Fairbairn et al., 2003). However, the effectiveness of such structures in subduing females is still uncertain, and only multivariate analyses of selection can disentangle their influence on the relative frequency of successful mating in natural populations.

image

Figure 1. Genital components of male Aquarius remigis. Scale bars indicate 0.5 mm. (a) Ventral view of retracted external genitalia: 7, 7th abdominal segment; 8, first genital segment (segment 8); Pg, pygophore (segment 9). (b) Lateral view of the distal abdominal segments and everted external genitalia in copulatory position. The proctiger (pr) is raised, and the phallus is emerging from the pygophore. Note the apical extension of the dorsal plate of the vesica (dp) in position to contact the female gonocoxae. (c) Ventral view of the vesica dissected from the genital capsule, showing the apical extension of the dorsal plate. Arrows indicate the landmarks used for measuring length of the dorsal plate. Panel A is adapted and reprinted with permission from Fairbairn et al. (2003). Panel C is adapted and reprinted with permission from Gallant & Fairbairn (1996).

Download figure to PowerPoint

In this study, we measure the influence of sexual selection acting through differential mating success on external and internal genital components of male A. remigis in six natural populations. Our goals are twofold. First, we assess if sexual selection for longer genitalia, which has previously been reported only for three populations from eastern North America (Preziosi & Fairbairn, 1996,2000; Ferguson & Fairbairn, 2000; Sih et al., 2002), is also pervasive in populations from California. Sexual selection is known to favour longer total length (= length of the somatic plus genital segments) in males from many California populations (Fairbairn & Preziosi, 1994,1996), and we specifically test the hypothesis that this apparent selection on overall body length can be attributed primarily to selection targeting genital length. Second, we use selection gradient analysis to identify which components of male external and internal genitalia are associated with increased male mating success. These analyses allow us to determine the adaptive significance of male genital morphology in the context of access to females, and hence to evaluate the influence of precopulatory sexual selection in the evolution of the singular genital morphology characteristic of males in the remigis group of species.

Study animal

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Study animal
  6. Sampling and measurements
  7. Measures of selection, statistical analyses
  8. Results
  9. Net sexual selection
  10. Direct sexual selection
  11. Relation between patterns of sexual selection on total length and genital length
  12. Discussion
  13. Sexual selection on male genital length and its relation to the selection on total length
  14. Sexual selection on the genital components, and the role of precopulatory sexual selection in shaping male genitalia
  15. Evolution of male genitalia: conclusions and perspectives
  16. Acknowledgments
  17. References

Aquarius remigis are semiaquatic bugs that are widely distributed on streams and small rivers throughout most of temperate and subtropical North America (Preziosi & Fairbairn, 1992; Gallant et al., 1993; Gallant & Fairbairn, 1996). In most populations, adults lack wings and dispersal among streams is rare (Calabrese, 1979; Fairbairn, 1986; Fairbairn & Desranleau, 1987). As a consequence, even geographically proximate natural populations show a high degree of genetic isolation (Blanckenhorn, 1991; Preziosi & Fairbairn, 1992).

When reproductively active, both sexes have an average of at least one successful mating per day (Preziosi & Fairbairn, 1996; Vermette & Fairbairn, 2002), although mating attempts are much more frequent (Weigensberg & Fairbairn, 1994). Successful copulations last on average 2–4 h but can extend to >12 h (Wilcox, 1984; Rubenstein, 1984; Weigensberg & Fairbairn, 1994; Campbell & Fairbairn, 2001; Vermette & Fairbairn, 2002). While in copula, the male rides on the back of the female, grasping her with his forelegs. The male maintains intromission throughout the entire duration of pairing, with sperm transfer occurring at the end of copulation just prior to withdrawal (Campbell & Fairbairn, 2001). The male then dismounts immediately, either voluntarily or, more commonly, as a result of a struggle by the female (Weigensberg & Fairbairn, 1994). In contrast to most Gerris species, (Arnqvist, 1997a), A. remigis males do not guard their mates by remaining mounted following copulation.

The morphology of male genitalia follows the general gerromorphan plan (Andersen, 1975,1982; Fairbairn et al., 2003). Genitalia are composed of a proximal cylindrical segment 8 (Fig. 1a) and a distal boat-shaped structure called the pygophore (segment 9, Fig. 1a) which is dorsally closed by a lid-like structure, the proctiger (segment 10, Fig. 1b). When deflated, the phallic organ lies inside the pygophore. Previous studies indicate that both segment 8 and the pygophore contribute to the interindividual variation in the total length of male genitalia and could facilitate intromission of the phallus (Fairbairn et al., 2003). However, Fairbairn et al. (2003) also postulated that elongation of the pygophore might be an indirect consequence of the selection favouring increased length or size of the phallus. As in other water striders, the intromittent organ consists of a proximal scletorized phallotheca and a distal endosoma which carries the vesica (Andersen, 1975,1982). However, the phallus of the remigis group is distinguished by the presence of a pronounced scletorized apical extension of the dorsal plate of the vesica (Fig. 1b, c) and a large, spiny ventral lobe that inflates during the copulation (Michel, 1961; Andersen, 1990; Damgaard & Cognato, 2003; Fairbairn et al., 2003). In A. remigis, the length of the apical extension of the dorsal plate is both genetically and phenotypically correlated with length of the external genitalia (Fairbairn et al., 2003; Fairbairn, unpublished data). Detailed observations of mating interactions have revealed that the apical extension of the dorsal plate forms the terminus of the inflating male phallus and makes first contact with the female gonocoxae (Fig. 1b, Fairbairn et al., 2003). Fairbairn et al. (2003) postulated that the dorsal plate of the vesica may therefore be the target of sexual selection acting on male genital length because it facilitates initial intromission, while the inflated ventral lobe may help in anchoring the phallus in female genital organs (Fairbairn et al., 2003).

Sampling and measurements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Study animal
  6. Sampling and measurements
  7. Measures of selection, statistical analyses
  8. Results
  9. Net sexual selection
  10. Direct sexual selection
  11. Relation between patterns of sexual selection on total length and genital length
  12. Discussion
  13. Sexual selection on male genital length and its relation to the selection on total length
  14. Sexual selection on the genital components, and the role of precopulatory sexual selection in shaping male genitalia
  15. Evolution of male genitalia: conclusions and perspectives
  16. Acknowledgments
  17. References

The water striders used for the present analyses were originally collected for assays of sexual selection on total body length in male A. remigis from Californian populations (Fairbairn & Preziosi, 1994,1996). These are cross-sectional samples obtained using hand nets to capture all adults found along a 50–100 m section of stream in a single 2–3 h sampling session. All samples were taken in fine weather (sunny and dry), between 1000 and 1600 h, when the water striders are most active (Gentile, 1998). Mark-recapture studies of A. remigis using the same capture methods have capture probabilities ranging from 76 to 99%, with values exceeding 90% typical for reproductive populations sampled in fine weather, as in the present study (Fairbairn, 1985; Preziosi & Fairbairn, 2000; Fairbairn, unpublished data). Thus, we are confident that mating and single animals are represented in their true proportions in our samples (Fairbairn & Preziosi, 1994). The collected water striders were separated according to mating status (single or paired), and mating pairs were retained in buckets until the pairs had separated and the male had fully deflated his phallus. All captured adults were then preserved in 70% ethanol for later measurement.

We randomly selected six of the original 12 populations for reanalysis (Table 1). All of these populations are sufficiently geographically separated to be considered genetically independent (Preziosi & Fairbairn, 1992; Fairbairn & Preziosi, 1994). Significant univariate selection (i.e. net selection) favouring longer males had been found in only two of the original 12 populations, neither of which were among the populations chosen for remeasurement (Table 1). However, multivariate analysis can be used to measure selection acting directly on each trait, with the effects of selection on other, correlated traits held constant (Lande & Arnold, 1983). The previous multivariate analysis, which included total length, wing morphology (winged or nonwinged), length of the mesofemur and width of the profemur, had revealed significant positive direct sexual selection on total length in seven of the original 12 populations, including four of the six populations chosen for remeasurement (Table 1). Thus, sexual selection tends to favour longer (i.e. larger) males in these populations. Significant nonlinear selection on total length had been detected in only one sample, and this sample was not included for remeasurement. Thus, our subsample of six populations is representative of the general trends seen in the original sample of 12 populations, with sexual selection indicated by significant positive multivariate linear selection coefficients for most but not all of the populations.

Table 1.  Population numbers*, locations, and original estimates* of linear (directional) selection on total length for the six Californian populations remeasured for the current study.
Population*LatitudeLongitudentnmSelection on total length*
β1Pβ3P
  1. β1 and β3 are univariate selection and multivariate selection coefficients. Total number of males (nt) and number of mating males (nm) refer to the numbers used for remeasurement (see text).

  2. *Population number, univariate (β1) and multivariate (β3) linear selection coefficients from Fairbairn & Preziosi (1994). Probabilities are 1-tailed, testing for positive β.

 139.5120150880.068ns0.783<0.005
 437.112252260.151ns0.656<0.01
 537.312264100.344ns1.402<0.005
 838.41209330−0.111ns−0.285ns
1040.31215012−0.367ns−0.498ns
1140.41227429−0.194ns0.643<0.005

We remeasured the preserved males using the protocols described in Preziosi & Fairbairn (1996), with the exception that measurements were taken directly from the specimens, rather than from photographic negatives. Males were placed in a standard position, ventral side up, a fixed distance from the lens of a binocular dissecting microscope, the image digitized, and linear measurements made from the digitized image using specialized software (Measurement TV© for samples 1–10, and SigmaScan Pro© for sample 11). We measured body components at 5×, external genital components at 20×, and the dorsal plate of the vesica at 40×. Repeatabilities of all measurements exceeded 0.97.

Body dimensions are robust to the effects of preservation in 70% ethanol (Brennan, 1993). However, to reduce the dimensionality of the multivariate analyses, and because we were nevertheless concerned that the male abdomens may have developed some curvature during storage, we did not include abdominal length in our measurement set. For each male we measured the length of the thorax, length of the mesofemur, width of the profemur (measured at midlength), total length of the genital segments, length of the first genital segment (segment 8, Fig. 1a), and length of the second genital segment (the pygophore, segment 9, Fig. 1a). Lengths of the body and genital components were taken at midline.

Following measurement of the above components of the whole animal, we dissected the vesica (Fig. 1b) with its attached, sclerotized dorsal plate (Fig. 1b, c) from the pygophore of each male, following the protocol described in Gallant & Fairbairn (1996) and Fairbairn et al. (2003). The vesicas were cleared and mounted on microscope slides, and the total length of the dorsal plate was digitized at 40× and measured from the distal tip of the apical extension of the plate to the proximal end of the vesicular capsule (Fig. 1c, Fairbairn et al., 2003).

Sample sizes for the remeasurements are given in Table 1. These are slightly lower than the sample sizes given in Fairbairn & Preziosi (1994) because only animals for which we could measure all of the body and genital components were included in the present analyses.

Measures of selection, statistical analyses

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Study animal
  6. Sampling and measurements
  7. Measures of selection, statistical analyses
  8. Results
  9. Net sexual selection
  10. Direct sexual selection
  11. Relation between patterns of sexual selection on total length and genital length
  12. Discussion
  13. Sexual selection on male genital length and its relation to the selection on total length
  14. Sexual selection on the genital components, and the role of precopulatory sexual selection in shaping male genitalia
  15. Evolution of male genitalia: conclusions and perspectives
  16. Acknowledgments
  17. References

We estimated standardized, multivariate selection gradients following the methods of Lande & Arnold (1983) and Fairbairn & Preziosi (1996). We included wing morphology (winged/nonwinged), and all of the measured body and leg components in the regression models, but we report new selection coefficients only for thorax and genital components. To preserve power for testing our hypotheses, we did not retest the hypotheses that sexual selection acts on wing and leg components, as these hypotheses had been evaluated from the original measurements of these samples (sexual selection tends to favour wingless males, while the patterns of sexual selection on leg measurement are generally weak and inconsistent among populations; Fairbairn & Preziosi, 1996). Thus, in our analyses, these variables are treated as control variables only.

To estimate the selection coefficients for our target traits, we standardized all trait values (zi) within each population, to a mean of 0 and a standard deviation of 1, and transformed mating success to relative fitness (winline image) by dividing individual mating success (0 for unpaired males and 1 for paired males) by the average male mating success of the population.

We examined the combined influence of both direct and indirect selection on genital components (i.e. the net selection) by calculating univariate selection coefficients. These are estimated from univariate regressions of relative fitness on standardized trait values. Univariate linear (β1) and nonlinear (γ1 = 2 β2) selection coefficients were obtained using the following models w′ = c + β1zi and w′ = c + inline imagezi + β2z2. Such selection coefficients measure the combined effects of direct selection acting on a trait and the indirect selection on this trait resulting from selection on correlated characters.

We also measured the strength of selection acting independently on each trait within each population (i.e. direct selection) from multivariate regressions of relative fitness on the standardized values of thorax length and genital length, plus the three control variables. Linear (β3) and nonlinear (γ2i = 2β4i) multivariate coefficients of selection were estimated from the models w′ =c + ∑β3izi and w′ = c + ∑βizi + ∑β4izinline image + ∑β5ijzizj. In this last model, the coefficient β5ij is a measure of the correlational selection between the traits i and j.

We tested the hypothesis that sexual selection on genital length targets specific genital components by computing linear multivariate coefficients as above, but replacing genital length with the lengths of segment 8, the pygophore and the length of the dorsal plate (hereafter ‘dorsal plate’). We did not compute the nonlinear selection coefficients in these cases since such models would have suffered from a reduced statistical power given the very large number of independent variables (n = 27).

All selection coefficients were estimated from least-squares regressions. The significance of the coefficients was, nevertheless, assessed by logistic regression since our measure of relative fitness was a binary variable violating general assumptions of the least square regression methods (Fairbairn & Preziosi, 1994,1996; Brodie et al., 1995). However, for one model (the model including all the genital traits for site 8), the logistic regression failed to find a solution. This occurs most commonly when there is a complete or quasi-complete separation of the data (i.e. when the dependent variable is almost perfectly predicted by the independent variables; Hosmer & Lemeshow, 1989; Pampel, 2000). In such a situation, the logistic coefficients and their standard errors tend to infinity (Hosmer & Lemeshow, 1989; Pampel, 2000). The quasi-complete separation for our data set is supported by the fact that the length of segment 8 and thorax length alone explained 73% of the variation on male mating success. Consequently, for this model we conducted a backward stepwise regression model initially including all variables and using an alpha of 0.05 for retention. Selection coefficients of the traits retained in the final model were considered significant.

Because several selection coefficients were estimated for each site, the probability values of these estimates should be corrected for multiple comparisons (Fairbairn & Preziosi, 1996; Preziosi & Fairbairn, 1996; Fairbairn & Reeve, 2001), although this practice has not been widely adopted for selection gradient analyses (Kingsolver et al., 2001). We thus indicate coefficients that were still significant after a sequential Bonferroni correction for all the estimated coefficients within each population.

Net sexual selection

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Study animal
  6. Sampling and measurements
  7. Measures of selection, statistical analyses
  8. Results
  9. Net sexual selection
  10. Direct sexual selection
  11. Relation between patterns of sexual selection on total length and genital length
  12. Discussion
  13. Sexual selection on male genital length and its relation to the selection on total length
  14. Sexual selection on the genital components, and the role of precopulatory sexual selection in shaping male genitalia
  15. Evolution of male genitalia: conclusions and perspectives
  16. Acknowledgments
  17. References

Our results agree with the pattern of selection detected for thorax length in eastern North American populations (Preziosi & Fairbairn, 1996,2000; Ferguson & Fairbairn, 2000) in suggesting weak net selection for reduced thorax length (β1 is negative in four populations; t-test: average β1 = −0.229, t5 = −2.44, P = 0.06; but is never significant after Bonferroni correction; Table 2). One of the nonlinear univariate selection coefficients is also negative and significantly different from zero before Bonferroni correction (site 1; Table 2). However, this does not reflect a general trend (only two of the γ1 for thorax length are negative while four are positive, Table 2; t-test: average γ1 = −0.009, t5 = −0.11, n.s.).

Table 2.  Univariate linear (β1) and nonlinear (γ1) selection coefficients for net sexual selection on male thorax length and the lengths of male genital components.
 Site 1Site 4Site 5Site 8Site 10Site 11
β1γ1β1γ1β1γ1β1γ1β1γ1β1γ1
  1. Significance was determined from logistic regressions using likelihood ratio chi-square tests. Sample sizes as in Table 1.

  2. Boldface values are significant after sequential Bonferroni corrections for all the estimated coefficients within sites.

  3. *P < 0.05; **P < 0.01; ***P < 0.001.

Thorax0.044−0.246**0.0560.226−0.5020.920−0.316*0.101−0.384−0.139−0.2720.019
Genital length0.532***−0.211−0.014−0.0870.1730.5270.571***0.023−0.342−0.046−0.307*−0.134
Segment 80.484***−0.080−0.0900.0110.026−0.2370.948***0.031−0.589*−0.107−0.110−0.258
Pygophore0.309***−0.1280.039−0.1980.2070.0690.0670.227−0.3540.119−0.2370.111
Dorsal plate0.105−0.137−0.074−0.2020.0010.704−0.3190.253−0.666*0.237−0.095−0.009

In contrast to the results for thorax length, but again consistent with studies of eastern populations (Preziosi & Fairbairn, 1996), net linear selection on genital length was strong, positive and highly significant in two of our populations (Table 2). However, we detected no significant net selection on genital length in the other four populations. The patterns of net selection on segment 8 were similar to those on total genital length (Table 2). These data demonstrate that the external genitalia are subject to strong, positive net directional selection and, as shown for genital length in previous studies, this is the strongest and most consistent pattern of sexual selection detected for morphological traits. However, we did not detect this in all populations: three of the net linear coefficients for these traits are positive but three are negative (t-test: average β1 = 0.102, t5 = 0.63, n.s., for genital length and average β1 = 0.112, t5 = 0.51, n.s., for segment 8 length). Net selection on the pygophore and dorsal plate was generally weaker than that on total genital length or the length of segment 8, with only one coefficient (pygophore in site 1) significant after Bonferroni correction (Table 2). Although this is evidence that net selection does favour increased pygophore length in some populations, neither pygophore nor dorsal plate appears to be under strong, consistent net selection (Table 2; four of the six β1 for pygophore are positive but t-test: average β1 = 0.005, t5 = 0.05, n.s.; only half of the β1 for dorsal plate are positive, t-test: average β1 = 0.0007, t5 = −0.17, n.s.).

Net nonlinear sexual selection on genital traits was not significant in any population (Table 2) and no overall trend was apparent (e.g. the average γ1varied from −0.107 to 0.155, t5 comprised between −0.17 and 0.51, n.s. in all cases).

Direct sexual selection

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Study animal
  6. Sampling and measurements
  7. Measures of selection, statistical analyses
  8. Results
  9. Net sexual selection
  10. Direct sexual selection
  11. Relation between patterns of sexual selection on total length and genital length
  12. Discussion
  13. Sexual selection on male genital length and its relation to the selection on total length
  14. Sexual selection on the genital components, and the role of precopulatory sexual selection in shaping male genitalia
  15. Evolution of male genitalia: conclusions and perspectives
  16. Acknowledgments
  17. References

The multivariate linear models including the two leg measures, wing morphology, and thorax and genital lengths are significant for four of the six populations (Table 3). However, for site 11, the overall model was significant but there was no significant selection on any of the traits of interest (the model significance in this case was due to selection acting on a control variable, i.e. wing morphology). Thus the multivariate analyses indicate significant selection on components of size for three of the six populations.

Table 3.  Multivariate linear (β3) and nonlinear (γ2) selection coefficients for sexual selection on male thorax and genital lengths. The coefficients are derived from multivariate models including leg measures and wing morphology as control variables (coefficients not reported; see text).
 Site 1Site 4Site 5Site 8Site 10Site 11
β3γ2β3γ2β3γ2β3γ2β3γ2β3γ2
  1. Significance was determined from logistic regressions using likelihood ratio chi-square tests. LR represents −2 log (likelihood ratio). Sample sizes as in Table 1.

  2. Boldface values are significant after sequential Bonferroni corrections for all the estimated coefficients within sites.

  3. *P < 0.05; **P < 0.01; ***P < 0.001.

Thorax length−0.256*0.3190.203−2.212*−0.7461.6931.358***0.5050.712−6.3060.3461.321
Genital length0.760***−0.3480.019−1.280*1.031*1.4700.878***−0.445−0.199−2.6060.014−0.191
Model
 r20.4890.6040.1120.4440.2060.4130.5790.7240.1600.3780.1870.299
 LR99.54***122.83***8.1032.04**11.40*22.8967.71***84.67***8.8320.8518.51**29.68*
 d.f.515515414515515515

For both thorax and genital length, the multivariate analyses indicate that the trends detected in the univariate analyses result from selection directly targeting these body components. For thorax length, the multivariate linear coefficient for site 8 is significant after Bonferroni correction and strongly negative (Table 3). However, as in the univariate analysis, this pattern of selection is not consistent among populations: selection tends to be stabilizing (but not significant) for site 4, and the linear coefficients are positive but not significant in sites 10 and 11 (Table 3). Thus, as in eastern populations, thorax length seems not to be subject to strong, consistent sexual selection. Nevertheless, when directional selection occurs, it favours decreasing rather than increasing thorax length.

The multivariate analysis of selection on genital length confirms the univariate results, showing strong, significant selection favouring longer genitalia in two populations (Table 3). Further, the statistical removal of the indirect effects of selection on other variables revealed a much stronger overall trend for positive selection on genital length: β3 is positive in five sites while β1 was positive in only three. Thus, in agreement with previous selection analyses (Preziosi & Fairbairn, 1996,2000; Ferguson & Fairbairn, 2000), the multivariate coefficients indicate that sexual selection tends to favour longer genital length, and that this pattern of selection can be very strong.

Although positive directional selection is clearly the only strong pattern of selection on genitalia evident in any of our populations, the nonlinear coefficients for genital length tend to be negative (five of six are negative; mean γ2 = −0.974, t5 = 2.373, P = 0.06), suggesting that the relationship between genital length and mating success (i.e. the fitness function) may be slightly curved, the slope decreasing at larger sizes (Phillips & Arnold, 1989). Site 4 shows this pattern most clearly (i.e. the linear coefficient is small, while the negative nonlinear coefficient is large and marginally nonsignificant), and in this population, correlational selection may account for the apparent curvature in the fitness surface. We detected significant correlational selection between genital length and both thorax and mesofemur lengths (β5 = 1.46, Likelihood ratio test: inline image = 12.00, P < 0.001 and −1.09, Likelihood ratio test: inline image = 11.40, P < 0.001). Specific interpretation of these coefficients is complex and must be made in the context of all of the coefficients in the full model (Phillips & Arnold, 1989), but from the perspective of the present study, the important message is that significant correlational selection indicates that the effect of genital length on mating success depends on the sizes of other body components. Thus, whether selection strongly favours increased genital length may depend at least in part on the size and/or allometry of body components of males in that population. Although the multivariate analysis supports such a mechanism for site 4, no other correlational selection coefficients were statistically significant after Bonferroni correction, in any population, and no general trends were evident. Larger samples, and hence higher statistical power, would be required to verify the curvature in the multivariate fitness surfaces, as well as adequately test the hypothesis that such curvature is associated with correlational selection.

Our analysis of multivariate linear selection on the genital components (segment 8, pygophore and dorsal plate of the vesisca) revealed direct linear selection on these traits only in populations in which selection also favoured longer overall genital length (Tables 3 and 4). Selection favouring increased length of segment 8 showed the same pattern of statistical significance as overall genital length. For pygophore, the pattern was similar for sites 1 and 5, but nonsignificant for site 8. Thus, with the latter exception, selection for overall longer genitalia is associated with similar selection on both external genital segments.

Table 4.  Multivariate linear selection coefficients (β3) for sexual selection on the lengths of male genital components. The coefficients are derived from multivariate models including leg measures, thorax length and wing morphology as control variables (coefficients not reported; see text).
 Site 1Site 4Site 5Site 8Site 10Site 11
  1. Significances were determined from logistic regressions using likelihood ratio chi-square tests but for site 8 significances were tested from a backward stepwise regression model (see text). LR represents −2 log (likelihood ratio). nt and nm as in Table 1.

  2. Boldface values are significant after sequential Bonferroni corrections for all the estimated coefficients within sites.

  3. *P < 0.05; ***P < 0.001.

  4. P < 0.05 after backward logistic regression.

Segment 80.446***−0.1780.747*1.140−0.340−0.157
Pygophore0.286***0.1130.710*0.038−0.1400.042
Dorsal plate0.114−0.1280.137−0.348−0.2760.175
Model
 r20.3840.1180.3070.9430.2130.221
 LR74.51***7.5515.56*90.12***10.1819.97*
 d.f.776377
 nt (nm)144 (86)46 (24)60 (9)77 (24)41 (11)60 (25)

In contrast to the results for external genital segments, we detected no significant directional selection on the length of the dorsal plate in any of the sites, and no general pattern (three positive and three negative coefficients, Table 4). Thus selection on the external genital segments is not associated with similar patterns of selection on the major sclerotized portion of the intromittent organ.

Relation between patterns of sexual selection on total length and genital length

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Study animal
  6. Sampling and measurements
  7. Measures of selection, statistical analyses
  8. Results
  9. Net sexual selection
  10. Direct sexual selection
  11. Relation between patterns of sexual selection on total length and genital length
  12. Discussion
  13. Sexual selection on male genital length and its relation to the selection on total length
  14. Sexual selection on the genital components, and the role of precopulatory sexual selection in shaping male genitalia
  15. Evolution of male genitalia: conclusions and perspectives
  16. Acknowledgments
  17. References

If selection on overall body length results from direct selection acting on genital length, we should find significant selection favouring longer total length only when selection also favours longer genitalia and vice versa. Our results are not consistent with these expectations. In sites 4 and 11, we found positive sexual selection on total length in the absence of significant directional selection on genital length (Tables 1 and 3). The reverse occurred in site 8, where sexual selection favoured longer genitalia (Table 3) while total length was not under sexual selection (Table 1). Thus, in three of our six populations, we have a mismatch between the pattern of selection on total and genital lengths. Comparison of Tables 1 and 3 suggests that the pattern of selection on total length depends on the combination of selection on both thorax and genital lengths, rather than on the latter alone. The patterns for sites 4 and 11 illustrate that selection can be significantly positive on total length if the coefficients are also positive for both thorax and genital lengths, even if these are individually nonsignificant. It may be that, in these populations, the target of selection is indeed overall body size. Conversely, the pattern in site 8 illustrates that, even if selection on genital length is strongly positive, negative selection on thorax length can result in nonsignificant or even negative selection on total length. These results suggest that the variation in patterns of sexual selection on total length derives from variation in sexual selection acting independently on somatic as well as genital components.

Sexual selection on male genital length and its relation to the selection on total length

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Study animal
  6. Sampling and measurements
  7. Measures of selection, statistical analyses
  8. Results
  9. Net sexual selection
  10. Direct sexual selection
  11. Relation between patterns of sexual selection on total length and genital length
  12. Discussion
  13. Sexual selection on male genital length and its relation to the selection on total length
  14. Sexual selection on the genital components, and the role of precopulatory sexual selection in shaping male genitalia
  15. Evolution of male genitalia: conclusions and perspectives
  16. Acknowledgments
  17. References

Our results confirm previous observations of sexual selection favouring longer genitalia in A. remigis (Preziosi & Fairbairn, 1996,2000; Ferguson & Fairbairn, 2000, Sih et al., 2002) and indicate that such selection is common throughout the species’ range. Significant sexual selection favoured increased genital length in two of six of our populations, and the same tendency, though not significant after Bonferroni correction, occurred in a third population. In agreement with previous studies (Preziosi & Fairbairn, 1996,2000; Ferguson & Fairbairn, 2000; Sih et al., 2002), we also identified genital length as the only component of total length subject to positive sexual selection; selection on somatic components being either nonsignificant or negative.

In spite of the weight of evidence for positive sexual selection on both genital and total length in A. remigis (Fairbairn, 1988; Sih & Krupa, 1992,1995; Kaitala & Dingle, 1993; Krupa & Sih, 1993; Fairbairn & Preziosi, 1994; Preziosi & Fairbairn, 1996,2000; Ferguson & Fairbairn, 2000; Sih et al., 2002), our results clearly demonstrate that the pattern of sexual selection on total length cannot be predicted from selection on genital length alone. The pattern of sexual selection on total length depends upon the direction and degree of selection on somatic components as well. This selection tends to be weaker and more variable than the selection on genitalia, but with an overall tendency for smaller soma to be associated with higher mating success (this study and Preziosi & Fairbairn, 1996,2000; Ferguson & Fairbairn, 2000; Sih et al., 2002). The functional basis of this selection is not well understood and several processes may be involved. Smaller males require less energy for movement and maintenance, and may be able to out-compete larger males in scramble competition for females when search costs are high (e.g. at low population densities) or food is scarce (Ghiselin, 1974; Krupa & Sih, 1993; Blanckenhorn et al., 1995). When food is limiting, smaller males also have longer reproductive lifespans and are able to persist in mating attempts longer than larger males (Blanckenhorn et al., 1995; Preziosi & Fairbairn, 1996). Males with smaller somatic components also remain in copula longer than larger males (Sih et al., 2002), but because copulation duration appears to be negatively associated with paternity success (fitness) in A. remigis (Vermette & Fairbairn, 2002), this may represent selection for larger rather than smaller soma. Whatever the functional mechanisms for small soma advantage in sexual competition, it is clear that sexual selection on total length at least partly reflects the net effects of the primarily negative selection on somatic components, particularly thorax length and the primarily positive selection on genital length, and that these processes balance differently in different populations. Both the rate at which males encounter females and the intensity of female resistance to mating are strongly influenced by biotic and abiotic aspects of the environment, including sex ratio, population density, food availability, predation risk and habitat structure (Weigensberg & Fairbairn, 1994; Sih & Krupa, 1992,1995; Krupa & Sih, 1993,1998; Blanckenhorn et al., 1995; Lauer et al., 1996). Thus, it should not be surprising to find different patterns of sexual selection in different populations. The frequent occurrence of strong, highly significant selection favouring long external genitalia, in spite of the diversity of environmental conditions in the different populations sampled (Fairbairn, 1988; Fairbairn & Preziosi, 1994; Ferguson & Fairbairn, 2000), argues strongly for its role in the evolution of prolonged genitalia in A. remigis.

Response to the antagonistic selection on genital and somatic components would be expected to be constrained by genetic correlations among components and, conversely, the pattern of opposing selection pressures would be expected to favour developmental and genetic uncoupling of the genital and somatic components. This pattern has indeed been detected in A. remigis: both full-sib and half-sib breeding experiments, on two different source populations, have revealed relatively weak genetic correlations between male genital and somatic components contrasting with much stronger genetic correlations among within these trait types (Preziosi & Roff, 1998; Fairbairn, unpublished data). Similarly, these studies reveal very low between-sex genetic correlations for genital length, a pattern that may have facilitated the evolution of the strong sexual dimorphism in genital size (male external genitalia average three times longer than comparable segments in females; Fairbairn, 1992; Preziosi & Fairbairn, 1996; Fairbairn, unpublished data).

Sexual selection on the genital components, and the role of precopulatory sexual selection in shaping male genitalia

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Study animal
  6. Sampling and measurements
  7. Measures of selection, statistical analyses
  8. Results
  9. Net sexual selection
  10. Direct sexual selection
  11. Relation between patterns of sexual selection on total length and genital length
  12. Discussion
  13. Sexual selection on male genital length and its relation to the selection on total length
  14. Sexual selection on the genital components, and the role of precopulatory sexual selection in shaping male genitalia
  15. Evolution of male genitalia: conclusions and perspectives
  16. Acknowledgments
  17. References

While congruent evidence indicates that males with long genitalia have higher mating success because they are better able to overcome vigorous female reluctance to mate (Sih et al., 2002; Fairbairn et al., 2003), it seems unlikely that this mechanism favours genital length per se (Preziosi & Fairbairn, 1996). By means of a detailed examination of the mating interactions, Fairbairn et al. (2003) identified several genital structures that may enable males to achieve intromission in spite of active female resistance. Such potential targets of sexual selection include the lengths of segment 8, the pygophore, and the apical extension of the dorsal plate. Fairbairn et al. (2003) attributed the elongation of the nonintromittent genital traits (segment 8 and the pygophore) to mechanical advantage, since these characters allow the genitalia to extend, flex, bend and probe the female gonocoxae and, therefore, presumably influence the outcome of the intromission attempts. The current patterns of sexual selection on male genital morphology are in agreement with this hypothesis.

Fairbairn et al. (2003) also postulated that the prolonged apical extension of the dorsal plate, which represents more than 40% of the length of the plate, may have evolved to facilitate intromission in the face of active female reluctance (Fairbairn et al., 2003). The apical extension makes first contact with the female gonocoxae and is used by males to probe them before the intromission of the phallus. Given that the dorsal plate occupies more than 75% of the length of the pygophore when the phallus is deflated, Fairbairn et al. (2003) suggested that sexual selection for increased length of the dorsal plate may favour increased length of the pygophore as a consequence of packing constraints. Under this hypothesis, selection on pygophore length should be an indirect consequence of direct selection for a longer dorsal plate. Our results do not support this hypothesis. We found no evidence that precopulatory sexual selection favours increased length of the dorsal plate, even in populations where net selection for increased pygophore length is significant. Thus contemporary patterns of selection do not support the hypothesis that male intromission success depends on length of the dorsal plate, or that the selection on pygophore length derives indirectly from the constraints of accommodating the deflated intromittent organ.

Our results support the hypothesis that sexual selection acting through differential mating success acts strongly on the size (and shape: see Fairbairn et al., 2003) of the external genitalia of male A. remigis. However, other processes must account for the evolution of the unique characteristics of the intromittent organ in the remigis group, as we found no evidence of sexual selection on the length of the dorsal plate. Three other hypotheses have been proposed to account for the evolution of components of the male genitalia that interact with female genital components during intromission. The first of these is the lock-and key hypothesis which postulates that selection for avoidance of hybridization has driven genital evolution so that male genital organs represent a species-specific key to female genitalia (the lock) (Dufour, 1844; Shapiro & Porter, 1989; Arnqvist, 1997b). To date, this hypothesis has not been well-supported by either comparative or intraspecific studies (Arnqvist, 1997b,1998). Although our data do not permit a comprehensive evaluation of the lock-and-key hypothesis in A. remigis (see Arnqvist, 1997b), our results disagree with some of its key predictions. Arnqvist (1997b) and Arnqvist et al. (1997) predicted that, under the lock-and-key hypothesis, the morphology of the intromittent organ should affect male mating success. In addition, except during speciation events, the overall selection on the phallic morphology acting through differential mating success should be stabilizing, since selection should favour the male genital configuration fitting the average female genital morphology (Alexander et al., 1997; Arnqvist, 1997b). Contrary to these predictions, we found evidence for neither a relationship between the dorsal plate and male mating success nor net stabilizing sexual selection. Thus, the lock-and-key hypothesis seems unlikely to explain the evolution of the elongated dorsal plate.

A second hypothesis for the evolution of the male intromittent organ postulates that rapid divergence of male genitalia is an indirect response to selection on genetically correlated characters (Mayr, 1963). This ‘pleiotropy hypothesis’ assumes that variation in genital morphology is largely selectively neutral (Arnqvist, 1997b) and has been criticized since it implies that genital traits are more sensitive to pleiotropic effects than other morphological characters (Hosken & Stockley, 2004). Evidence, nevertheless, suggests that pleiotropic effects may have contributed to the morphological diversification of male genitalia of Gerris species (Arnqvist & Thornhill, 1998). Indeed, in Gerris incognitus, the morphology of the vesical sclerites, which is highly variable within the genus, is affected by the selection acting on genetically and phenotypically correlated characters (Arnqvist et al., 1997; Arnqvist & Thornhill, 1998). However, A. remigis seems to differ from Gerris species in this. Indeed, while the length of the dorsal plate is both phenotypically and genetically correlated with some body length components (Fairbairn et al., 2003; range of genetic correlations from full-sib experiment with 80 full-sib families and 726 male offspring: 0.05–0.48; Fairbairn, unpublished data) suggesting that pleiotropic gene effects might account for the evolution of this character, the current patterns of selection for mating success do not provide support for the maintenance of the elongated dorsal plate by pleiotropy. However, contrary to the pleiotropy hypothesis, we did not detect indirect selection (i.e. net selection) on this intromittent component, and in the one population where it approached significance (site 10) it was actually favouring reduction of the length of the dorsal plate, a pattern opposite to both the morphological trend (prolongation) and the selection on correlated genital traits.

The third hypothesis for the evolution of male intromittent organs, the ‘post-copulatory sexual selection hypothesis’, postulates that genital evolution results from the influence of genital morphology on male fertilization success (Eberhard, 1985,1997,1998,2001; Arnqvist, 1998). While our data do not allow us to directly test this hypothesis, several lines of evidence indicate ample opportunity for post-intromission sexual selection in A. remigis. In particular, the proportion of variation in paternity success that can be explained by mating success alone is relatively low (r2 ≤ 0.36), which suggests that post-intromission processes such as cryptic female choice and sperm competition might account for the residual variance in male fertilization success (Vermette & Fairbairn, 2002). In addition, the unusually long copulations that characterize A. remigis provide ample opportunity for such processes to occur (Campbell & Fairbairn, 2001). Finally, a significant influence of the shape of the apical extension of the dorsal plate on mating duration has been detected (Bertin, Fairbairn and Vermette, unpublished data). Post-intromission sexual selection may thus play a significant role in shaping the male intromittent organ in A. remigis as it does in other water strider species (Arnqvist & Danielsson, 1999; Danielsson & Askenmo, 1999). Further experiments evaluating the influence of genital morphology on paternity success will be necessary to elucidate the full functional significance of the singular morphology of male genitalia in A. remigis, but the results of the present study suggest that both precopulatory and post-copulatory processes are likely to be important.

Evolution of male genitalia: conclusions and perspectives

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Study animal
  6. Sampling and measurements
  7. Measures of selection, statistical analyses
  8. Results
  9. Net sexual selection
  10. Direct sexual selection
  11. Relation between patterns of sexual selection on total length and genital length
  12. Discussion
  13. Sexual selection on male genital length and its relation to the selection on total length
  14. Sexual selection on the genital components, and the role of precopulatory sexual selection in shaping male genitalia
  15. Evolution of male genitalia: conclusions and perspectives
  16. Acknowledgments
  17. References

Our results join the growing number of empirical studies demonstrating that genital function of males goes beyond sperm transfer to include contributions to the relative mating and/or paternity success of competing males (e.g. Mason, 1980; Arnqvist, 1989; Eberhard, 1993a,b,1996,2001; Preziosi & Fairbairn, 1996; Arnqvist et al., 1997; Arnqvist & Danielsson, 1999; Danielsson & Askenmo, 1999; Ferguson & Fairbairn, 2000; Sih et al., 2002; House & Simmons, 2003; Takami, 2003; Yang & Wang, 2004). Male genitalia thus function as both primary sexual traits and as secondary sexual traits whose form and function have evolved in response to sexual selection. However, few studies have evaluated the influence of the morphology of male intromittent organs on mating success (but see Goulson, 1993). In our study, we did not find evidence for such a relationship, in spite of the apparent functional significance of the apical extension of the dorsal plate. Given the relatively low power of our analyses (Hersch & Phillips, 2004) to detect stabilizing selection acting directly on individual genital traits, it would be premature to conclude that the phallic morphology of A. remigis has not been influenced by precopulatory sexual selection. Nevertheless, our evidence does suggest that precopulatory processes exert a much greater effect on the evolution of external genital morphology than on the morphology of the intromittent organ itself.

Other studies have also demonstrated significant effects of nonintromittent genital structures on male mating ability (Mason, 1980; Arnqvist et al., 1997; Sih et al., 2002). Such genital structures, generally related to the grasping of females, seem to be common in insects (Eberhard, 2004a). In water striders of the genus Gerris, intersexual conflicts arising from different optimal mating rates in males and females are widespread and have led to the correlated evolution of the morphology of males and females (Arnqvist & Rowe, 1995,2002a,b; Rowe & Arnqvist, 2002). The structures involved in this antagonistic coevolution comprise the pregenital and genital segments in males and the abdominal spines in females (Arnqvist & Rowe, 1995,2002a). Long and curved genitalia allow males to better overcome female resistance whereas long and elevated spines improve ability of females to resist male mating attempts (Arnqvist & Rowe, 2002a). A similar sexual conflict over mating decisions occurs in A. remigis (Weigensberg & Fairbairn, 1994; Sih & Krupa, 1995; Watson et al., 1998; Sih et al., 2002), but in this species there is no evidence of a similar pattern of antagonistic coevolution of male and female pregenital and genital segments (Fairbairn et al., 2003). Nevertheless, our study does support the hypothesis that male genitalia have evolved in response to sexual selection for the ability to overcome female reluctance to mate, and hence, that sexual conflicts over mating rates have contributed to the rapid divergence of male genitalia in the Gerridae.

Given that differences in the optimal mating rates of males and females are probably pervasive (Trivers, 1972; Parker, 1979; Rice, 2000), one might expect sexual conflict over mating to have played a significant role in driving the diversity of male genitalia in many taxa with internal fertilization. However, the validity of this hypothesis has been questioned (Eberhard, 1985,1997, 1998,2004a,b; Eberhard et al., 1998; Hosken & Stockley, 2004), and comparative studies indicate that sexual conflict is not sufficient to explain patterns of genitalic diversity in many (and perhaps most) taxa (Eberhard, 2004a,b). As suggested by our results, selection may act on the genitalia of males through a variety of mechanisms, influencing both mating and fertilization success, and it seems unlikely that any one mechanism will provide a general explanation for their diversification (Simmons, 2001). We agree with Hosken & Stockley (2004) that efforts can more profitably be directed toward evaluating the relative contributions of the mechanisms involved. By quantifying the relationship between mating success and male genital morphology in natural populations, our study provides strong evidence for sexual selection on genital components through precopulatory sexual selection driven, at least partly, by sexual conflict over mating decisions. Our results suggest that direct quantification of precopulatory sexual selection on genitalia is a necessary complement to studies of processes occurring during intromission (e.g. cryptic female choice, copulatory courtship, and sperm competition) and to the comparative approach taken by Eberhard (2004a,b)). Only through such a comprehensive approach can we hope to discern the true adaptive significance of male genitalic diversity.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Study animal
  6. Sampling and measurements
  7. Measures of selection, statistical analyses
  8. Results
  9. Net sexual selection
  10. Direct sexual selection
  11. Relation between patterns of sexual selection on total length and genital length
  12. Discussion
  13. Sexual selection on male genital length and its relation to the selection on total length
  14. Sexual selection on the genital components, and the role of precopulatory sexual selection in shaping male genitalia
  15. Evolution of male genitalia: conclusions and perspectives
  16. Acknowledgments
  17. References
  • Alexander, R., Marshall, D. & Cooley, J. 1997. Evolutionary perspectives on insect mating. In: The Evolution of Mating Systems in Insects and Arachnids (J.Choe & B.Crespi, eds). pp. 131. Cambridge University Press, Cambridge.
  • Andersen, N.M. 1975. The Limnogonus and Neogerris of the old world with character analysis and a reclassification of the Gerrinae (Hemiptera: Gerridae). Entomol. Scand. 7 (Suppl.): 196.
  • Andersen, N.M. 1982. The Semiaquatic Bugs (Hemiptera: Gerromorpha): Phylogeny, Adaptations, Biogeography and Classification. Scandinavian Science Press, Klampenborg, Denmark.
  • Andersen, N.M. 1990. Phylogeny and taxonomy of water striders, genus Aquarius Schellenberg (Insecta, Hemiptera, Gerridae), with a new species from Australia. Seenstrupia 16: 3781.
  • Arnaud, L., Haubruge, E. Gage, M.J.G. 2001. Morphology of Tribolium castaneum male genitalia and its possible role in sperm competition and cryptic female choice. Belg. J. Zool. 131: 111115.
  • Arnqvist, G. 1989. Sexual selection in a water strider – the function, mechanism of selection and heritability of a male grasping apparatus. Oikos 56: 344350.
  • Arnqvist, G. 1997a. The evolution of water strider mating systems: causes and consequences of sexual conflicts. In: The Evolution of Mating Systems in Insects and Arachnids (J.Choe & B.Crespi, eds), pp. 146163. Cambridge University Press, Cambridge, New York.
  • Arnqvist, G. 1997b. The evolution of animal genitalia: distinguishing between hypotheses by single species studies. Biol. J. Linn. Soc. 60: 365379.
  • Arnqvist, G. 1998. Comparative evidence for the evolution of genitalia by sexual selection. Nature 393: 784786.
  • Arnqvist, G. & Danielsson, I. 1999. Postmating sexual selection: the effects of male body size and recovery period on paternity and egg production rate in a water strider. Behav. Ecol. 10: 358365.
  • Arnqvist, G. & Rowe, L. 1995. Sexual conflict and arms races between the sexes – a morphological adaptation for control of mating in a female insect. Proc. R. Soc. Lond B Biol. Sci. 261: 123127.
  • Arnqvist, G. & Rowe, L. 2002a. Correlated evolution of male and female morphologies in water striders. Evolution 56: 936947.
  • Arnqvist, G. & Rowe, L. 2002b. Antagonistic coevolution between the sexes in a group of insects. Nature 415: 787789.
  • Arnqvist, G. & Thornhill, R. 1998. Evolution of animal genitalia: patterns of phenotypic and genotypic variation and condition dependence of genital and non-genital morphology in water strider (Heteroptera: Gerridae: Insecta). Genet. Res. 71: 193212.
  • Arnqvist, G., Thornhill, R. & Rowe, L. 1997. Evolution of animal genitalia: morphological correlates of fitness components in a water strider. J. Evol. Biol. 10: 613640.
  • Artiss, T. 2001. Structure and function of male genitalia in Libellula, Ladona and Plathemis (Anisoptera: Libellulidae). Odonatologica 30: 1327.
  • Blanckenhorn, W.U. 1991. Life-history differences in adjacent water strider populations: phenotypic plasticity or heritable responses to stream temperature? Evolution 45: 15201525.
  • Blanckenhorn, W.U., Preziosi, R.F. & Fairbairn, D.J. 1995. Time and energy constraints and the evolution of sexual size dimorphism - to eat or to mate. Evol. Ecol. 9: 369381.
  • Brennan, J.M. 1993. Clinal variation in morphology in eastern populations of the stream-dwelling waterstrider, Aquarius remigis (Hemiptera: Gerridae). M. Sc. Thesis, Concordia University, Montreal, Canada.
  • Brodie, E.D., Moore, A.J. & Janzen, F.J. 1995. Visualizing and quantifying natural selection. Trends Ecol. Evol. 10: 313318.
  • Calabrese, D.M. 1979. Pterygomorphism in 10 nearctic species of gerris. Am. Midl. Nat. 101: 6168.
  • Campbell, V. & Fairbairn, D.J. 2001. Prolonged copulation and the internal dynamics of sperm transfer in the water strider Aquarius remigis. Can. J. Zool. 79: 18011812.
  • Cordoba-Aguilar, A. 2000. Evolucion y diversidad de la morfologia de los genitales masculinos en insectos. Folia Entomol. Mex. 110: 95111.
  • Damgaard, J., Andersen, N.M. & Sperling, F.A.H. 2000. Phylogeny of the water strider genus Aquarius Schellenberg (Heteroptera: Gerridae) based on nuclear and mitochondrial DNA sequences and morphology. Insect. Syst. Evol. 31: 7190.
  • Damgaard, J. & Cognato, A. 2003. Sources of character conflict in a clade of water striders (Heteroptera: Gerridae). Cladistics 19: 512526.
  • Danielsson, I. & Askenmo, C. 1999. Male genital traits and mating interval affect male fertilization success in the water strider Gerris lacustris. Behav. Ecol. Sociobiol. 46: 149156.
  • Darwin, C. 1871. The Descent of Man and Selection in Relation to Sex. Murray, London.
  • Dufour, L. 1844. Anatomie générale des Diptères. Ann. Sci. Nat. 1: 244264.
  • Eberhard, W.G. 1985. Sexual Selection and Animal Genitalia. Harvard University Press, Harvard.
  • Eberhard, W.G. 1993a. Evaluating models of sexual selection – genitalia as a test-case. Am. Nat. 142: 564571.
  • Eberhard, W.G. 1993b. Copulatory Courtship and Morphology of Genitalic Coupling in 7 Phyllophaga Species (Coleoptera, Melolonthidae). J. Nat.l Hist. 27: 683717.
  • Eberhard, W.G. 1996. Female Control: Sexual Selection by Cryptic Female Choice. Princeton University Press, Princeton, New Jersey.
  • Eberhard, W.G. 1997. Sexual selection by cryptic female choice in insects and arachnids. In: The Evolution of Mating Systems in Insects (J.Choe & B.Crespi, eds), pp. 3257. Cambridge University Press, Cambridge.
  • Eberhard, W.G. 1998. Female roles in sperm competition. In: Sperm Competition and Sexual Selection (T.Birkhead & A.Moller, eds), pp. 91116. Academic Press, London.
  • Eberhard, W.G. 2001. Species-specific genitalic copulatory courtship in sepsid flies (Diptera, Sepsidae, Microsepsis) and theories of genitalic evolution. Evolution 55: 93102.
  • Eberhard, W.G. 2004a. Male–female conflict and genitalia: failure to confirm predictions in insects and spiders. Biol. Rev. 79: 121186.
  • Eberhard, W.G. 2004b. Rapid divergent evolution of sexual morphology: comparative tests of antagonistic coevolution and traditional female choice. Evolution 58: 19471970.
  • Eberhard, W.G., Huber, B.A., Rodriguez, R.L., Briceno, R.D., Salas, I. & Rodriquez, V. 1998. One size fits all? Relationships between the size and degree of variation in genitalia and other body parts in twenty species of insects and spiders. Evolution 52: 415431.
  • Fairbairn, D.J. 1985. Comparative ecology of Gerris remigis (Hemiptera: Gerridae) in 2 Habitats – a Paradox of Habitat Choice. Can. J. Zool. 63: 25942603.
  • Fairbairn, D.J. 1986. Does alary dimorphism imply dispersal dimorphism in the waterstriderGerris remigis? Ecol. Entomol. 11: 355368.
  • Fairbairn, D.J. 1988. Sexual selection for homogamy in the gerridae: an extension of Ridley's comparative approach. Evolution 42: 12121222.
  • Fairbairn, D.J. 1992. The origins of allometry: size and shape polymorphism in the common waterstrider, Gerris remigis Say. Biol. J. Linn. Soc. 45: 167186.
  • Fairbairn, D.J. & Desranleau, L. 1987. Flight threshold, wing muscle histolysis, and alary polymorphism – correlated traits for dispersal tendency in the Gerridae. Ecol. Entomol. 12: 1324.
  • Fairbairn, D.J. & Preziosi, R.F. 1994. Sexual selection and the evolution of allometry for sexual size dimorphism in the water strider, Aquarius-Remigis. Am. Nat. 144: 101118.
  • Fairbairn, D.J. & Preziosi, R.F. 1996. Sexual selection and the evolution of sexual size dimorphism in the water strider, Aquarius remigis. Evolution 50: 15491559.
  • Fairbairn, D.J. & Reeve, J.P. 2001. Natural selection. In: Evolutionary Ecology. Concepts and Case Studies. (C. W.Fox, D. A.Roff & D. J.Fairbairn, eds), pp. 2943. Oxford University Press, Oxford.
  • Fairbairn, D.J., Vermette, R., Kapoor, N.N. & Zahiri, N. 2003. Functional morphology of sexually selected genitalia in the water strider Aquarius remigis. Can. J. Zool. 81: 400413.
  • Ferguson, I.M. & Fairbairn, D.J. 2000. Sex-specific selection and sexual size dimorphism in the waterstrider Aquarius remigis. J. Evol. Biol. 13: 160170.
  • Gallant, S.L. & Fairbairn, D.J. 1996. A new species of Aquarius from the southeastern United States, with electrophoretic analysis of the clade containing Gerris, Limnoporus, and Aquarius (Hemiptera: Gerridae). Ann. Entomol. Soc. Am. 89: 637644.
  • Gallant, S.L., Preziosi, R.F. & Fairbairn, D.J. 1993. Clinical variation in eastern populations of the waterstrider Aquarius Remigis– Gradual intergradation or discontinuity. Evolution 47: 957964.
  • Gentile, G. 1998. Daily patterns of oviposition mediate sexual conflict in a polygynandrous mating system. M.Sc.Thesis, Concordia University, Montreal, Canada.
  • Ghiselin, M. 1974. The Economy of Nature and the Evolution of Sex. University of California Press, Berkeley, CA USA.
  • Goulson, D. 1993. Variation in the genitalia of the butterfly Maniola jurtina (Lepidoptera: Satyrinae). Zool. J. Lin. Soc. 107: 6571.
  • Grant, V. 1995. Sexual selection in plants – pros and cons. Proc. Nat. Acad. Sci. USA 92: 12471250.
  • Hersch, E.I. & Phillips, P.C. 2004. Power and potential bias in field studies of natural selection. Evolution 58: 479485.
  • Hosken, D.J. & Stockley, P. 2004. Sexual selection and genital evolution. Trends Ecol. Evol. 19: 8793.
  • Hosmer, D. & Lemeshow, S. 1989. Applied Logistic Regression. Wiley, New York.
  • House, C.M. & Simmons, L.W. 2003. Genital morphology and fertilization success in the dung beetle Onthophagus taurus: an example of sexually selected male genitalia. Proc. R. Soc. Lond B Biol. Sci. 270: 447455.
  • Kaitala, A. & Dingle, H. 1993. Wing dimorphism, territoriality and mating frequency of the waterstrider Aquarius-remigis (say). Annales Zoologici Fennici. 30: 163168.
  • Kingsolver, J.G., Hoekstra, J.M., Berrigan, D., Vignieri, S.N., Hill, C.E., Hoang, A., Gibert, P. & Beerli, P. 2001. The strength of phenotypic selection in natural populations. Am. Nat. 157: 245261.
  • Krupa, J.J. & Sih, A. 1993. Experimental studies on water strider mating dynamics – spatial variation in density and sex-ratio. Behav. Ecol. Sociobiol. 33: 107120.
  • Krupa, J.J. & Sih, A. 1998. Fishing spiders, green sunfish, and a stream-dwelling water strider: male–female conflict and prey responses to single versus multiple predator environments. Oecologia 117: 258265.
  • Lande, R. & Arnold, S.J. 1983. The measurement of selection on correlated characters. Evolution 37: 12101226.
  • Lauer, M.J., Sih, A. & Krupa, J.J. 1996. Male density, female density and inter-sexual conflict in a stream-dwelling insect. Anim. Behav. 52: 929939.
  • Mason, L.G. 1980. Sexual selection and the evolution of pair-bonding in soldier beetles. Evolution 34: 174180.
  • Mayr, E. 1963. Animal Species and Evolution. Harvard University Press, Cambridge, MA.
  • Michel, F. 1961. The taxonomic value of male genitalia of the genus Gerris Fabricus(Hemiptera: Heteroptera: Gerridae). M. Sc. Thesis, Oregon State University, Corvallis, Oregon.
  • Pampel, F. 2000. Logistic Regression: A Primer. Sage Publications, Thousand Oaks, California.
  • Parker, G.A. 1979. Sexual selection and sexual conflict. In: Selection Selection and Reproductive Competition in Insects (M. S.Blum & N. A.Blum, eds), pp. 123166. Academic Press, New York.
  • Phillips, P.C. & Arnold, S.J. 1989. Visualizing multivariate selection. Evolution 43: 12091222.
  • Preziosi, R.F. & Fairbairn, D.J. 1992. Genetic population-structure and levels of gene flow in the stream dwelling waterstrider, Aquarius (=Gerris) remigis (Hemiptera, Gerridae). Evolution 46: 430444.
  • Preziosi, R.F. & Fairbairn, D.J. 1996. Sexual size dimorphism and selection in the wild in the waterstrider Aquarius remigis: body size, components of body size and male mating success. J. Evol. Biol. 9: 317336.
  • Preziosi, R.F. & Fairbairn, D.J. 2000. Lifetime selection on adult body size and components of body size in a waterstrider: Opposing selection and maintenance of sexual size dimorphism. Evolution 54: 558566.
  • Preziosi, R.F. & Roff, D.A. 1998. Evidence of genetic isolation between sexually monomorphic and sexually dimorphic traits in the water strider Aquarius remigis. Heredity 81: 9299.
  • Rice, W.R. 2000. Dangerous liaisons. Proc. Nat. Acad. Sci. USA 97: 1295312955.
  • Rowe, L. & Arnqvist, G. 2002. Sexually antagonistic coevolution in a mating system: combining experimental and comparative approaches to address evolutionary processes. Evolution 56: 754767.
  • Rubenstein, D.I. 1984. Resource acquisition and alternative mating strategies in water striders. Am. Zool. 24: 345353.
  • Shapiro, A.M. & Porter, A.H. 1989. The lock-and-key hypothesis: evolutionary and biosystematic interpretation of insect genitalia. Ann. Rev. Entomol. 34: 231245.
  • Sih, A. & Krupa, J.J. 1992. Predation risk, food-deprivation and nonrandom mating by size in the stream water strider, Aquarius-remigis. Behav. Ecol. Sociobiol. 31: 5156.
  • Sih, A. & Krupa, J.J. 1995. Interacting effects of predation risk and male and female density on male–female conflicts and mating dynamics of stream water striders. Behav. Ecol. 6: 316325.
  • Sih, A., Lauer, M. & Krupa, J.J. 2002. Path analysis and the relative importance of male-female conflict, female choice and male-male competition in water striders. Anim. Behav. 63: 10791089.
  • Simmons, L. 2001. Sperm Competition and its Evolutionary Consequences in the Insects. Princeton University Press, Princeton.
  • Sirot, L.K. 2003. The evolution of insect mating structures through sexual selection. Fla. Entomol. 86: 124133.
  • Takami, Y. 2003. Experimental analysis of the effect of genital morphology on insemination success in the ground beetle Carabus insulicola (Coleoptera Carabidae). Ethol. Ecol. Evol. 15: 5161.
  • Trivers, R.L. 1972. Parental investment and sexual selection. In: Sexual Selection and the Descent of Man (B.Campbell, ed.), pp. 136179. Heinemann, Chicago.
  • Vermette, R. & Fairbairn, D.J. 2002. How well do mating frequency and duration predict paternity success in the polygynandrous water strider Aquarius remigis? Evolution 56: 18081820.
  • Watson, P.J., Arnqvist, G. & Stallmann, R.R. 1998. Sexual conflict and the energetic costs of mating and mate choice in water striders. Am. Nat. 151: 4658.
  • Weigensberg, I. & Fairbairn, D.J. 1994. Conflicts of interest between the sexes – a study of mating interactions in a semiaquatic bug. Anim. Behav. 48: 893901.
  • Weigensberg, I. & Fairbairn, D.J. 1996. The sexual arms race and phenotypic correlates of mating success in the waterstrider, Aquarius remigis (Hemiptera: Gerridae). J. Ins. Behav. 9: 307319.
  • Wilcox, R.S. 1984. Male copulatory guarding enhances female foraging in a water strider. Behav. Ecol. Sociobiol. 15: 171174.
  • Yang, L.H. & Wang, Q. 2004. Precopulation sexual selection in Nysius huttoni White (Heteroptera: Lygaeidae) in relation to morphometric traits. J. Ins. Behav. 17: 695707.